From: "LIZHAOXIN1 [李照鑫]" <LIZHAOXIN1@kingsoft.com>
To: "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
"quintela@redhat.com" <quintela@redhat.com>,
"dgilbert@redhat.com" <dgilbert@redhat.com>,
"berrange@redhat.com" <berrange@redhat.com>
Cc: "LIZHAOXIN1 [李照鑫]" <LIZHAOXIN1@kingsoft.com>
Subject: Re: [PATCH v2] migration/rdma: Use huge page register VM memory
Date: Wed, 16 Jun 2021 13:02:20 +0000 [thread overview]
Message-ID: <145b82f8ce5b4db2a605e2ee1480a909@kingsoft.com> (raw)
> When using libvirt for RDMA live migration, if the VM memory is too large,
> it will take a lot of time to deregister the VM at the source side, resulting
> in a long downtime (VM 64G, deregister vm time is about 400ms).
>
> Although the VM's memory uses 2M huge pages, the MLNX driver still uses 4K
> pages for pin memory, as well as for unpin.
> So we use huge pages to skip the process of pin memory and unpin memory to reduce downtime.
>
The test environment:
kernel: linux-5.12
MLNX: ConnectX-4 LX
libvirt command:
virsh migrate --live --p2p --persistent --copy-storage-inc --listen-address \
0.0.0.0 --rdma-pin-all --migrateuri rdma://192.168.0.2 [VM] qemu+tcp://192.168.0.2/system
> ---
> v2
> - Add page_size in struct RDMALocalBlock
> - Use page_size to determine whether VM uses huge page
> ---
>
> Signed-off-by: lizhaoxin <lizhaoxin1@kingsoft.com>
>
> diff --git a/migration/rdma.c b/migration/rdma.c index 1cdb4561f3..703816ebc7 100644
> --- a/migration/rdma.c
> +++ b/migration/rdma.c
> @@ -215,6 +215,7 @@ typedef struct RDMALocalBlock {
> uint64_t remote_host_addr; /* remote virtual address */
> uint64_t offset;
> uint64_t length;
> + uint64_t page_size;
> struct ibv_mr **pmr; /* MRs for chunk-level registration */
> struct ibv_mr *mr; /* MR for non-chunk-level registration */
> uint32_t *remote_keys; /* rkeys for chunk-level registration */
> @@ -565,7 +566,8 @@ static inline uint8_t *ram_chunk_end(const RDMALocalBlock *rdma_ram_block,
>
> static int rdma_add_block(RDMAContext *rdma, const char *block_name,
> void *host_addr,
> - ram_addr_t block_offset, uint64_t length)
> + ram_addr_t block_offset, uint64_t length,
> + uint64_t page_size)
> {
> RDMALocalBlocks *local = &rdma->local_ram_blocks;
> RDMALocalBlock *block;
> @@ -595,6 +597,7 @@ static int rdma_add_block(RDMAContext *rdma, const char *block_name,
> block->local_host_addr = host_addr;
> block->offset = block_offset;
> block->length = length;
> + block->page_size = page_size;
> block->index = local->nb_blocks;
> block->src_index = ~0U; /* Filled in by the receipt of the block list */
> block->nb_chunks = ram_chunk_index(host_addr, host_addr + length) + 1UL; @@ -634,7 +637,8 @@ static int qemu_rdma_init_one_block(RAMBlock *rb, void *opaque)
> void *host_addr = qemu_ram_get_host_addr(rb);
> ram_addr_t block_offset = qemu_ram_get_offset(rb);
> ram_addr_t length = qemu_ram_get_used_length(rb);
> - return rdma_add_block(opaque, block_name, host_addr, block_offset, length);
> + ram_addr_t page_size = qemu_ram_pagesize(rb);
> + return rdma_add_block(opaque, block_name, host_addr, block_offset,
> + length, page_size);
> }
>
> /*
> @@ -1123,13 +1127,25 @@ static int qemu_rdma_reg_whole_ram_blocks(RDMAContext *rdma)
> RDMALocalBlocks *local = &rdma->local_ram_blocks;
>
> for (i = 0; i < local->nb_blocks; i++) {
> - local->block[i].mr =
> - ibv_reg_mr(rdma->pd,
> - local->block[i].local_host_addr,
> - local->block[i].length,
> - IBV_ACCESS_LOCAL_WRITE |
> - IBV_ACCESS_REMOTE_WRITE
> - );
> + if (local->block[i].page_size != qemu_real_host_page_size) {
> + local->block[i].mr =
> + ibv_reg_mr(rdma->pd,
> + local->block[i].local_host_addr,
> + local->block[i].length,
> + IBV_ACCESS_LOCAL_WRITE |
> + IBV_ACCESS_REMOTE_WRITE |
> + IBV_ACCESS_ON_DEMAND |
> + IBV_ACCESS_HUGETLB
> + );
> + } else {
> + local->block[i].mr =
> + ibv_reg_mr(rdma->pd,
> + local->block[i].local_host_addr,
> + local->block[i].length,
> + IBV_ACCESS_LOCAL_WRITE |
> + IBV_ACCESS_REMOTE_WRITE
> + );
> + }
> if (!local->block[i].mr) {
> perror("Failed to register local dest ram block!\n");
> break;
Hello everyone, please give me some advice.
Thanks.
lizhaoxin1
next reply other threads:[~2021-06-16 13:19 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-16 13:02 LIZHAOXIN1 [李照鑫] [this message]
-- strict thread matches above, loose matches on Subject: below --
2021-06-10 15:39 [PATCH v2] migration/rdma: Use huge page register VM memory LIZHAOXIN1 [李照鑫]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=145b82f8ce5b4db2a605e2ee1480a909@kingsoft.com \
--to=lizhaoxin1@kingsoft.com \
--cc=berrange@redhat.com \
--cc=dgilbert@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).